Amir Skandani a, J. Arul Clement a, Stephanie Tristram-Nagle b, M. Ravi Shankar a, *
a Swanson School of Engineering, 3700 O'Hara Street, University of Pittsburgh, Pittsburgh, PA 15261, USAb Biological Physics Group, Physics Department, 5000 Forbes Avenue, Carnegie Mellon University, Pittsburgh, PA 15213, USA
a r t i c l e i n f o
Article history:Received 23 August 2017Received in revised form20 October 2017Accepted 24 October 2017Available online 5 November 2017
The aliphatic spacer length connected to the rigid mesogenic cores is shown to control the structuralorganization and mechanical actuation in response to light-stimulus in azobenzene-functionalized liquidcrystalline polymers networks. The spacer lengths in the mesogenic host (non-photochromic) and in thephoto-active azobenzene-functionalized cross linker are parametrically varied to create monodomainliquid crystalline samples. A suite of thermomechanical, photomechanical and structural characterizationis used to characterize the binary co-polymers. The photomechanical responses are compared bycalculating a figure-of-merit e Photocompliance (Cp). This parameter, Cp, which characterizes the in-cremental strain generated in response to unit intensity of irradiation (W/cm2), is found to correlatestrongly with the Sxray order parameter and the D-spacings. We show that compact (smaller D) andhigher ordering of the mesogens that result in copolymers with longer flexible spacers manifest greaterphotomechanical activity.
Liquid crystalline polymers (LCP) embedded with anisotropy atthe molecular level offer programmable materials with responsestunable to an array of stimuli, including heat [1,2], light [3,4], sol-vent [5,6] andmechanical deformation [7]. Utilizing light to actuateLCP that are embedded with photochromic moieties such as azo-benzene has attracted significant attention. This is motivated by theopportunity for enabling new system designs that use light as acontactless energy source, which can be irradiated from largestand-off distances with spatiotemporal modulation of the in-tensity and polarization, across length-scales. This has been utilizedin photoresponsive cilia [8], customizable topographies with fric-tion that can be tuned with light at the micrometer-scale, creationof photomechanical robots that can manipulate objects at themillimeter length-scale [9], photomobile “walkers” andmotors thatuse multiplexed (uv and >500 nm visible light) irradiation togenerate macroscopic motion [10], swimmers that can operateunder photomechanically-coupled hydrodynamic conditions [11]and digital actuators that can be toggled between discrete states
using light [3].Research efforts have examined two broad classes of azo-
functionalized LCP e i) softer liquid crystal elastomers manifest-ing substantial actuation strains (>10%) albeit generating subduedlevels of stresses [12,13] and, ii) glassy, often main-chain, azo-functionalized liquid crystalline networks (ALCN) that are capableof manifesting actuation stresses approaching the MPa-scale, butcapable of modest strains (<10%) [4,14e16]. ALCN are useful fortheir applications as actuators in engineered applications due totheir substantial photomechanical work potential and stiffness.Typically, irradiation of ALCN under ambient conditions, well belowtheir glass transition temperature, leads to generation of significantwork densities, which can be harnessed in an array of geometries[3]. A number of studies have focused on characterizing thestructure-property linkages underpinning the photoactuation inALCN. A promising class of ALCN was identified, which is charac-terized by host mesogens that are not photoresponsive, amongstwhich mesogenic photoresponsive moieties (e.g. azobenzene-based) are dispersed [17]. Photoinduced strains in such co-polymers are a function of the macromolecular structure andcomposition in addition to the intensity and wavelength of theirradiating light. Studies on the effects of cross-linking and con-centration of photoactive azobenzene molecular switches on thephoto-responsiveness of the ALCN have shown that the degree of
isomerization decreased with increasing the cross-link density,while addition of azo-based mesogen content up to 20 mol%monotonically increased the photomechanical work output [17].The improvement in response with increasing azo content yieldsdiminishing returns when it approaches ~35mol% due to progres-sive loss of nematic order at these compositions [16,18]. Increasingthe cross-linking density by increasing the degree of polymeriza-tion resulted in higher stiffness, but decreased the magnitude ofphotostrains [19].
Following irradiation with UV light, photostrains in ALCN aregenerated as a result of isomerization of rod-like trans states, intothe bent cis form. The associated molecular contraction is trans-mitted to the crosslinked polymer network, which manifests thephotomechanical response in bulk. The time-scales associated withthe trans-cis isomerization are on the order of picoseconds. How-ever, the bulk response which is mediated by the transduction ofthe isomerization-induced distortion amongst the polymernetwork is slower by several orders of magnitude [20]. The kineticsof the isomerization and ultimately the photoresponsiveness of theactuation is underpinned by the stability of the liquid crystal sys-tem. This controls the dynamics of the photo-strain accumulationand relaxation, which is in turn a function of the structure andcomposition [21]. The photoresponsiveness of ALCNs is alsomodified by the gradient in the absorption of the light through thethickness of the sample. The absorption of the light by the photo-chromic moieties leads to a non-linear intensity profile, which inturn manifests a gradient in the photostrains. This results in thecharacteristic bending of the samples towards the light source [22].
The transduction of photonic energy into mechanical deforma-tion in azobenzene-functionalized systems is mediated by the free-volume available for the trans-cis isomerization reaction [23,24],which is a function of the photochromic moiety, its conformationand its point of attachment to the polymeric chains [24]. In ALCN,the population of the cis and trans-forms of the azobenzene reachesa photostationary state in response to irradiation. The photosta-tionary state is a statistical equilibrium, which is characterized byphotomechanical strain saturation. The evolution towards this stateis a function of the irradiation (intensity, wavelength and polari-zation), which determines the forward and reverse isomerizationand reorientation of the photochromic switches. The dynamics of
O
OOO
O
NN
OO
O
n
Azo Crosslinkers n=4
RM257
O
OO
RM82
O
O
Fig. 1. Structure of host mesogens (RM257: 3C spacer and RM82: 6C spacer), which are indivvarious aliphatic chain lengths. Samples denoted by RM3C, RM3C-6C, RM3-8C, RM3C-10containing 6C, 8C and 10C aliphatic chain lengths, respectively. RM6C, RM6C-6C, RM6-8C, Rrespectively. nþ2 is the number of carbon atoms in the flexible chain of the azobenzene cr
such photoresponse, which generates the strain, is also associatedwith the generation of free volume as well as modification of themechanical properties of the material [25]. The evolution of prop-erties and the kinetics of the photoisomerization are intrinsicallyfunctions of the molecular flexibility in the vicinity of the photo-chromic moiety. This motivates us to examine whether modifyingthe molecular flexibility in the vicinity of the azobenzene via thechoice of aliphatic spacer lengths can modulate and magnify theresponsiveness of crosslinked ALCN.
Here, we first examine how the length of the aliphatic spacerconnected to the azobenzene-rigid core in the crosslinker and thatof the host modifies the ordering, mesogenic packing and thethermomechanical properties of the ALCN. Then, by defining afigure-of-merit e Photocompliance, we determine correlationsbetween the photomechanical response and structural character-istics as a function of the monomer structure. We examine two hostmesogens with various lengths of aliphatic spacers connected tothe mesogenic core: 2-Methyl-1,4-phenylene-bis[4[3(acryloyloxy)propyloxy]benzoate] (RM257; 3-carbon aliphatic spacer) and 2-Methyl-1,4-phenylene-bis[4[6(acryloyloxy)hexyloxy]benzoate](RM82; 6-carbon aliphatic spacer). Each of these was copoly-merized with azobenzene-functionalized mesogens of variousflexible spacer lengths (see Fig.1 for structures and Fig.1 caption forthe abbreviation that refers to the various compositions). Main-chain azobenzene-based diacrylate-functionalized mesogens weresynthesized and their concentration was fixed at 10 mol % for allsamples. By measuring the thermal, mechanical and photo-responsiveness of the resulting polymerized films, we delineate therole of the structure of the aliphatic spacer in the azobenzene-crosslinker and the host mesogens on the photomechanical prop-erties. In contrast to merely observing the geometric change inresponse to irradiation, we account for the gradients in the ab-sorption through the thickness and calculate the saturated photo-strains. To compare the responses of the various materials, weutilize photocompliance (saturation photostrain per intensity ofirradiation) as the distinguishing metric. Characterizing the struc-tural, thermomechanical and photomechanical properties via acombinatorial examination of the compositions reveals a frame-work for magnifying the photomechanical response.
O
OO O
O
O O
O
n,6,8
O
OO
O
O
idually copolymerized with azobenzene-functionalized cross linkers that each containC refer to copolymers of RM257 with azobenzene-functionalized (Azo) cross linkersM6C-10C refer to copolymers of RM82 with azobenzene cross linkers of n¼ 4, 6, and 8,oss linker.
A. Skandani et al. / Polymer 133 (2017) 30e3932
2. Materials and methods
2.1. Synthesis of 4,4-dihydroxyazobenzene (2)
A solution of p-aminophenol (10.0 g, 91.64 mmol) in diluted 1MHCl solution (200 mL) was cooled to 0 �C by immersion in an icebath. Next, NaNO2 (9.34 g, 109.8 mmol) was dissolved in 150 ml ofwater and added drop-wise to the solution of p-aminophenol.Then, to the diazotized solution, 200 ml of pre-cooled methanolwas added to. The resulting solution mixture was stirred for 1 h.Subsequently, phenol (8.62 g, 91.64 mmol) of dissolved in 65 ml of3 M aqueous sodium hydroxide was added drop-wise and reactionmixture was stirred at room temperature for 2 h. Methanol wasremoved by evaporation and concentrated HCl was added to adjustpH < 5, and the resulting precipitate was collected by filtration andwashed out all of the acid with water (See Scheme 1).
A solution of 4,4-dihydroxyazobenzene (5 g, 1 equiv), K2CO3 (3equiv), and 6-chlorohexan-1-ol (2.5 equiv) were dissolved in DMF(60 mL). A trace amount of KI was added, and the reaction mixturewas heated to 140 �C for 12 h. The reactionmixturewas poured intowater (1000 mL) and dried under vacuum so that the precipitatedsolid can be collected. Anhydrous MgSO4 was used to dry theorganic layer after extracting crude product with ethyl acetate. Theresulting solid was fed to next step without further purificationafter evaporation of the solvent.
2.3. General procedure for acrylation
A solution of O-alkylation (3.0 g, 1 equiv), trimethylamine (2.5equiv) and a trace amount of hydroquinone were dissolved in THF(60 mL) at 0 �C. The acryloyl chloride (3 equiv) was added dropwiseand the reaction mixture was stirred in two stages for 1 h and 12 hat 0 �C and room temperature respectively. The reaction wasquenched by adding 20 ml aqueous solution of NaHCO3 and stirredfor 30 min. The organic phase was separated and washed withwater. After complete removal of the solvent purification processwas preformed using column chromatography (silica gel, eluent:chloroform). The remaining solid was recrystallized from ethanolyielding a yellowish powder.
2.4. Polarized optical microscopy and differential scanningcalorimetry of monomers
Table 1 summarizes different monomer mixture compositionsthat were tested using a Perkin Elmer Diamond differential scan-ning calorimeter (DSC) and hot stage polarized optical microscopy(POM) to obtain the characteristic temperature of each mixture (i.e.melting temperature and nematic to isotopic phase trans-formation). Differential scanning calorimetry tests were performedwith a heating mode with 5 �C/min and under nitrogenenvironment.
2.5. Monodomain liquid crystal polymer samples
All six compositions outlined in Table 1 were added with Irga-cure 784, (photoinitiator,1 wt % of the total reactants) and heated at(130 �C) in the isotropic phase. The molten mixture was theninfiltrated into the parallel rubbed polyimide-coated glass cells of50 mm gap via capillarity. The parallel anchoring conditions oneither side of the cell ensure the creation of monodomain align-ment parallel to the rubbing direction. The filled cells were slowlycooled down at � 1� C/min to the nematic phase while Tp/TNI ratiowas fixed at 0.95, to attempt to maintain a comparable order
kers with spacers of various aliphatic spacer lengths.
Table 1The monomer composition and TNI temperature of different samples.
parameter across all samples [26,27]. Tp is the polymerizationtemperature and TNI is the nematic-isotropic transition tempera-ture of the monomer mixture. Edmund MI-150 high-intensityilluminator was utilized to irradiate the samples with a cutoff filter(l � 420 nm) in order to photopolymerize the samples for 1 h. Cellswere opened using scalpel to reveal the free-standing, mono-domain ALCN films.
2.6. Wide-angle X-ray scattering (WAXS)
WAXS data were collected using the Rigaku rotating anodeRUH3R. Samples were prepared by cutting a narrow strip(2 � 10 mm) from the thin polymer sheets, in the direction parallelto the rubbing direction (nematic director). These strips were fixedonto the axle of a precision motor (AirProducts, Inc.,) with double-stick tape, so that 3 mm extended out from the axle. This extensionwas X-rayed at 90� to the nematic director while the 2 mm widesample was rapidly rotated in the X-ray beam from -5 to 5�. Cu Karadiation with l ¼ 1.5418 Å was used; beam size was 0.5 � 0.5 mm,focused with a Xenocs Fox2D focusing collimator. 2D images werecollected on a Rigaku Mercury CCD detector in 5 min dezingeredscans. The sample-to-CCD distance was ~100 mm, in three separateruns, calibrated with Teflon powder (D ¼ 4.898 Å).
In order to carry out the WAXS data analysis, the images werefirst rotated counterclockwise by 90� and flipped horizontally. TheWAXS scattering was then primarily on the equator in the right-hand quadrant. This orientation is required for data analysis witha MATLAB liquid-crystal fitting program [28]. This program de-termines Sxray, a chain order parameter, similar to an NMR orderparameter [29]. More details about the WAXS data collection andanalysis are given in Supplementary Materials.
2.7. Thermomechanical characterization
A Perkin Elmer 8000 dynamic mechanical analysis (DMA) wasutilized to measure the mechanical properties of the monodomainsamples along the nematic director, including storage and lossmoduli and tan d. The DMA experiments were carried out in straincontrol mode using a tension fixture with frequency of 1 Hz andtemperature sweep from room temperature to 150 �C. DMA sam-ples were prepared in the form of 25 mm � 1 mm rectangles(longer axes parallel to the nematic director) and mounted with0.1 N of pre-stretching force to eliminate possible slack. Resultswere recorded and analyzed as a function of the composition of thefilms.
2.8. Photomechanical characterization
The absorption coefficient of each sample was measured usingCRAIC QDI 2010 UV-VIS-NIR instrument in 200e800 nm range.Rectangular samples with 25 mm length and 1 mm width wereexcised from the original films with the longer axes parallel to thenematic director. Samples were then clamped from one end andirradiated uniformly with 30 mW/cm [2] of un-polarized 375 nm
UV source until the photomechanical strain saturates and pro-gressive bending with continued irradiation stopped. A prior flatsample characteristically bends towards the light and develops acurvature due to the photostrains. Under these irradiation in-tensities, negligible photothermal effects are expected and theactuation essentially occurs photochemically under nominallyambient conditions [30]. The resultant deformations in the sampleswere recorded with a USB camera positioned at a right angle and inplane with the sample to observe the development of the photo-strained geometry. The recorded images were then sliced to 1frame per 2 s andwere analyzed using a home-built MatLab code toextract the evolution of the curvature of the sample throughout theexperiment. The saturation value of the curvature is used to char-acterize the photomechanical response.
3. Results and discussion
Fig. 2 shows the DSC peaks of different monomermixtures alongwith their hot stage polarized optical microscopy images betweeneach peak. It can be seen that by increasing the aliphatic chainlength the nematic to isotropic transformation temperature (TNI)decreases while the presence of azobenzene crosslinker reducesthe melting temperature irrespective of its structure. The effect ofthe spacer length on the TNI and the melting temperature isconsistent with prior studies [31,32]. The absorption coefficients(Napierian) of the polymer film compositions from Table 1 areshown in Fig. 3 as a function of wavelength ranging from 200 nm to800 nm. We note a marginal decline in the absorption coefficientsof samples with increasing carbon chain length except, RM3C-10C.Table 2 lists the coefficient at 375 nm at which we examine thephotomechanical responses.
Representative thermo-mechanical responses of the variouscomposition films is shown in Fig. 4. The reported storage modulus(E0) from DMA tests (Fig. 4a) is a measure of the sample's stiffnessalong the nematic director. It can be seen that E0 at room temper-ature is primarily determined by the host mesogen, where higherstiffness is observed for the shorter spacer length (RM257)(Table 2). Both storagemodulus and loss modulus (E') (Fig. 4b) dropsharply during glass transition before reaching a plateau at highertemperatures. Similar behavior of azo-LCNs have been reportedelsewhere [19]. From tand plot of the samples (Fig. 4c), a relativelybroad glass transition can be observed for all of the samples, whichis typical [14,15]. The cross-link density was calculated usingequation (1) where Thigh and E
0high are at temperature 50 �C higher
than Tg and its corresponding storage modulus respectively [19]:
Ve ¼ E0high
.3RThigh
(1)
The thermomechanical properties are shown in Table 2 for thevarious compositions.
The wide-angle X-ray scattering data are shown in Fig. 5. Theseare 2D CCD images of slightly more than one quadrant of the entireimage, where white pixels indicate higher X-ray scattering in-tensity. The X-ray beam is covered by a semi-transparent
Fig. 2. DSC curves of different compositions in the form of powder mixture beforepolymerization along with their POM images a) RM257 and b) RM82. Insets illustratesthe POM images (crossed polarizers) measured at the corresponding temperatures.
A. Skandani et al. / Polymer 133 (2017) 30e3934
molybdenum strip in the lower left corner. Absence of lamellarscattering close to the beam confirms that there is no smecticordering in these samples. Most of the scattered intensity is locatedon the meridian; the width and decrease of this intensity in anazimuthal direction are used to quantitate Sxray, an order parametersimilar to an NMR order parameter (Ref. 33, 34). As shown in Fig. 5,a bright, compact reflection near 1.5�1 in qz is indicative of awell-ordered polymer chain; this occurred in all samples. Sxray orderparameters and D-spacings are summarized in Table 2. The con-sequences of these order parameters and D-spacing values will bediscussed below. In addition to the major, bright meridionalreflection resulting from interactions of the main polymer chainbackbone, there is a very weak reflection at ~0.8 �1, with a D-spacing of ~8.4 Šin all of the samples. The origin of this small
reflection is unknown.The photomechanical response of the samples under UV irra-
diation is shown in Fig. 6. Instead of merely comparing the finalcurvatures of the sample, we seek to abstract a photomechanicalfigure-of-merit e the photocompliance (Cp). Cp is defined as thecontractile strain accumulated by an element of the material alongthe nematic director when exposed to irradiation of unit intensity(W/cm [2]). Normalizing the strain generated with respect to theintensity allows the results to hold broader applicability, whileallowing a one-to-one comparison between the various materials.When a nominally flat film is irradiated, the absorption through thethickness generates an intensity gradient through the film (Fig. 7).Strains generated within the monodomain film at any given dis-tance x from the surface scale with the intensity I(x):
IðxÞ ¼ Ioe�mx (2)
m is the absorption coefficient as a function of the composition andIo is the incident irradiation of the light at the surface of the sample.Note that the mole% of azobenzene is fixed across all compositionsin this study.
The photostrain value at depth x is determined by Cp value ofthe composition, where:
εph ¼ Cp IðxÞ (3)
The mechanical deformation manifesting bending that leads tocurvature of radius r can be described by a strain field εb, which canbe written as [15]:
εb ¼ xrþ c (4)
where, h is the thickness of the sample and x2½0,h] and c is thestretching strain. Imaging of the sample is used to measure r in ourexperiments by characterizing the deformed geometry.
The stress sy, which manifests the bending is given by Ref. [15]:
sy ¼ E0�xrþ c� εph
�; (5)
where E0 is the storage modulus (from DMA) along the nematicdirector.
Enforcing equilibrium for force (Z h
0sydx ¼ 0) and moments
(Z h
0xsydx ¼ 0Þ, the following expression was derived using
Mathematica. The measured radius of curvature (r) and the cur-vature (k) are:
r ¼ 1=k ¼ � h3m2emh
6IoCp�mhþ mh emh þ 2� 2emh
� (6)
Hence, the Cp value as a function of compositions can be backcalculated from an imaging-based measurement of the final cur-vature k of the irradiated sample using:
Cp ¼ � kh3m2emh
6Io�mhþ mh emh þ 2� 2emh
� (7)
Equation (7) and the saturated curvatures in Fig. 6 for thevarious compositions were used to calculate the absolute value ofCp (Table 2). Some overarching trends are observed from theseresults. Longer flexible spacer lengths in both the host and theazobenzene cross linkers result in larger values of Cp. Thus, for unitphoton flux, RM6C-10C produces the greatest strains (Cp ¼ 0.3),while RM 3C-6C produces the smallest (Cp ¼ 0.13). The difference
Fig. 3. Absorption coefficient of polymerized samples measured between 200 and 800 nm at room temperature.
Table 2Photo- and thermo-mechanical properties of the samples.
between the two compositions is by a factor greater than two.Indeed, the glass transition also shows a decline with increasingspacer length, which is consistent with prior reports [33]. Anotherfactor that contributes to the magnified response in the RM82 (6C)systems is the deeper penetration of light in this system, in com-parison to the RM257 (3C) [34]. This can magnify the isomerizationof the azobenzene and result in greater photoresponsiveness.
The WAXS data indicate while all systems are predominantlyordered, differences in the order parameter are observed, eventhough the ratio of the polymerization temperature with respect tothe nematic-isotropic transition of the monomer mixture was fixedat 0.95 for all compositions. The Sxray values for the RM3C samplesaveraged to 0.40, which shows that they are not as well-ordered asthe RM6C samples (average Sxray ¼ 0.50). These order parametersare similar to the highest order parameter of 0.47 determined forLCN-16.6, which is 98.4% RM82 (equivalent to RM6C series in thiswork) in the literature [35]. In addition, all of the RM6C sampleshad smaller D-spacings (average ¼ 4.46 Å) than the RM3C samples(average ¼ 4.50 Å), indicating a tighter packing. In an effort tounderstand how chain ordering and D-spacing are correlated withphotoresponse (Cp), two additional comparisons were made. WhenCp is plotted vs. Sxray order parameter as shown in Fig. 8 (a), a
dependence of Cp on Sxray is shown; i.e., the photoresponse in-creases in more ordered samples. When comparing D-spacing andCp as shown in Fig. 8 (b), the photoresponse is larger when thesample has a smaller D-spacing. These results suggest that themaximal photoresponse occurs when the polymer sample is well-ordered and also compact. In the case of RM6C-10C, the strain inan infinitesimal element of the material at the surface, which isdirectly exposed to the incident irradiation, can be calculated as~ Cp Io ¼ 0.9%. This is a significant value of strain in a glass ALCNwith a storagemodulus of ~1.75GPa that is irradiatedwith amodestintensity of light. It has been suggested that photoresponsiveness ismainly controlled by the free volume distribution around thephotochromic moiety, which is different from its usual randomdistribution. During photoisomerization, there can be a redistri-bution of free volume [24]. It is perhaps that the isomerizationproceeds most efficiently with a well-ordered, compact startingconfiguration. Starting from this configuration, trans-cis photo-isomerization generates disordering and concomitant photome-chanical strains most efficiently.
Instructive ideas can be drawn from analogous strain generationvia thermally-driven order-disorder transitions in liquid crystallinepolymers. In glassy LCP that are thermally-responsive, modest
Fig. 4. Thermomechanical properties of ALCN as a function of monomer structure. DMA experiments were conducted parallel to the nematic director with 1HZ frequency, a) Elasticmodulus (E0), b) Loss modulus (E'), c) tan(d).
A. Skandani et al. / Polymer 133 (2017) 30e3936
changes (5%) in the order parameter account for the strain gener-ation. In such systems, no transition to the isotropic state occurs,although the responsiveness is affected by the length of thealiphatic spacer that is attached to the mesogenic core; the longeraliphatic spacers correlate with greater responsiveness [27]. It hasbeen shown that longer aliphatic spacer length correlates withlower glass transition temperature (Tg) [33]. Indeed, photome-chanical strain generation in glassy ALCN well below Tg does notrequire the same level of segmental mobility that underpins ther-momechanical strain generation. But, higher segmental mobilitythat emerges with longer flexible spacers can be expected tomagnify photomechanical response. On the other hand, it is knownthat the length of the flexible spacer affects the stability of thenematic state [36], which determines the process windowavailablefor polymerizing the mesogens to inherit the molecular ordering.Utilizing longer flexible spacers in the azobenzene-functionalized
cross-linkers can adversely modify the achievement of nematicordering in mixtures created with the host mesogens (e.g. RM257/RM82). This constrains the ability to indefinitely magnify thephotoresponse by using longer flexible spacers.
Segmental mobility of the polymer network has an effect onthe isomerization efficiency. It has also been shown that chang-ing the aliphatic chain length in the spacer can affect the effi-ciency of the trans-cis isomerization as a function of the polymernetwork in temperatures below their Tg [37e39]. Flexibility ofthe molecular chains enabled by spacers attached to the meso-genic cores offers the necessary rotational flexibility required forreorientation [40]. This is clearly illustrated in Fig. 9, where the Tgof the various samples is plotted against Cp. It is clear that greaterthe aliphatic chain length, lower is the Tg and greater is thephotoresponse. It is noteworthy that this photoresponse alsogoes hand-in-hand with smaller D-spacing and higher Sxray when
Fig. 5. 2D CCDWAXS images at grazing angle-of-incidence to ALCN oriented 90� to thealignment direction.
0
0.02
0.04
0.06
0.08
0.1
0.12
0.14
0 20 40 60 80 100 120
3c10c 6c6c 6c8c6c10c 3c8c 3c6cRM 3C-10C RM 6C-8C
RM 3C-8CRM 6C-6C
RM 6C-10C RM 3C-8C RM 3C-6C
Curvature(/mm)
Irradia on Time (s)
Ligh
tON
Zero CurvatureIni ally Flat Sample
Photo-StrainSaturatedSamplewithFinite Curvature
Ligh
tOFF
Fig. 6. Photomechanical bending experiments of samples under 30 mW/cm [2] ofunpolarized 375 nm UV irradiation.
Fig. 7. Schematic of the geometric evolution, which was used to derive a figure-of-merit, photocompliance Cp, which is the normalized strain accumulated per unit in-tensity (W/cm [2]) of incident irradiation.
Fig. 8. (a) Photoresponse (Cp) vs. Sxray. Sxray errors were obtained by measuring twoseparate pieces of the same polymer film. (b) Photoresponse (Cp) vs. WAXS D-spacing.D-spacing errors were obtained by measuring two separate pieces of the same polymerfilm.
A. Skandani et al. / Polymer 133 (2017) 30e39 37
the aliphatic spacer length is made longer.An important caveat emerges from the observation that the
reorientation during isomerization of azobenzene involves pre-dominantly localized molecular distortions involving molecularsegments in the vicinity of the photochromic switch, which limitsrole for the structure to the spacer itself [41]. While the kinetics of
the trans to cis isomerization plays a deterministic role in theresponsiveness, so does the kinetics of the back-reaction. The cis-trans back isomerization rate is lower for azobenzene embedded inpolymeric hosts with longer spacers [37]. This effect can enable thegeneration of a greater concentration of the cis form in a photo-stationary state, which results in greater photomechanical strains.While remaining cognizant of the complex role of the molecularstructure on the photoresponse, we also note the observed corre-lations between the photomechanical output and the material'sbulk mechanical properties. The monomer structure (host and theazobenzene cross linkers) controls the resulting storage/lossmoduli and cross-link densities. The photoresponse that isobserved is known to be highly correlated with the mechanicalproperties of the matrix [42]. If the interplay of the molecularstructure with the modulus and photoinduced strain can bedelineated, a framework can emerge for modulating the propertycombinations. In particular, if the photostrain generation can beenhanced, while retaining a high modulus, material systems canemerge that offer work densities that allow their utilization in
Fig. 9. Decline in Tg with increasing spacer lengths correlates with greater Cp.
A. Skandani et al. / Polymer 133 (2017) 30e3938
practical light-responsive actuators. Such material property en-hancements can allow light-sensitive actuation to complementthermally-responsive counterparts fabricated from liquid crystal-line systems. For example, modifying the flexible spacer has beenshown to double the photocompliance, even when the mole % ofazobenzene is held constant. Thus, engineering the local molecularenvironment magnifies the resulting photomechanical actuation.
4. Conclusions
An exploration of the photomechanical response as a function ofthe aliphatic spacer length attached to the mesogenic code inazobenzene-functionalized liquid crystalline polymer networkswas undertaken. The spacer length connected to the azobenzenecore as well as that of the non-photochromic host mesogens werevaried. The azobenzene concentration was held constant amongthe binary combinatorial co-polymers and ratio of the temperatureof polymerization to that of the nematic-isotropic transition of themonomer mixes was fixed. Isolating the effect of the monomerstructure on the structural characteristics resulted in identifying aroute of magnifying the photostrain generation by incorporatinggreater flexibility in the networks using longer aliphatic spacers.Additionally, a figure of merit e photocompliance was calculatedfrom experiments that encapsulated the photostrain magnitudealong the nematic director per unit irradiation intensity. The pho-tocompliance was found to strongly correlate with the higher orderparameters, lower Tg and smaller D-spacing e compact, flexible,well-ordered nematic networks produce the greatest photostrains.This offers a motif for rationally designing GPa-stiffness glassy,macromolecular networks with orderedmolecular switches, wherethe photomechanical strain generation can be magnified, evenunder ambient conditions and modest irradiation intensities.
Acknowledgement
MRS acknowledges support from the Air Force Office of Scien-tific Research (FA9550-14-1-0229) and the National Science Foun-dation (#1435489). STN acknowledges the Physics department atCarnegie Mellon University for support for this work.
Appendix A. Supplementary data
Supplementary data related to this article can be found athttps://doi.org/10.1016/j.polymer.2017.10.050.
References
[1] T.J. White, M.E. McConney, T.J. Bunning, Dynamic color in stimuli-responsivecholesteric liquid crystals, J. Mater. Chem. 20 (44) (2010) 9832e9847.
[2] J. Küpfer, H. Finkelmann, Liquid crystal elastomers: influence of the orienta-tional distribution of the crosslinks on the phase behaviour and reorientationprocesses, Macromol. Chem. Phys. 195 (4) (1994) 1353e1367.
[3] M.R. Shankar, M.L. Smith, V.P. Tondiglia, K.M. Lee, M.E. McConney, D.H. Wang,L.-S. Tan, T.J. White, Contactless, photoinitiated snap-through in azobenzene-functionalized polymers, Proc. Natl. Acad. Sci. 110 (47) (2013) 18792e18797.
[4] Y. Yu, M. Nakano, T. Ikeda, Photomechanics: directed bending of a polymerfilm by light, Nature 425 (6954) (2003), 145e145.
[5] L.T. de Haan, J.M. Verjans, D.J. Broer, C.W. Bastiaansen, A.P. Schenning, Hu-midity-responsive liquid crystalline polymer actuators with an asymmetry inthe molecular trigger that bend, fold, and curl, J. Am. Chem. Soc. 136 (30)(2014) 10585e10588.
[6] T. Kamal, S.-Y. Park, A liquid crystal polymer based single layer chemo-responsive actuator, Chem. Commun. 50 (16) (2014) 2030e2033.
[7] H. Aoyama, Y. Yamazaki, N. Matsuura, H. Mada, S. Kobayashi, Alignment ofliquid crystals on the stretched polymer films, Mol. Cryst. Liq. Cryst. 72 (4)(1981) 127e132.
[8] H. Yang, G. Ye, X. Wang, P. Keller, Micron-sized liquid crystalline elastomeractuators, Soft Matter 7 (3) (2011) 815e823.
[9] F. Cheng, R. Yin, Y. Zhang, C.-C. Yen, Y. Yu, Fully plastic microrobots whichmanipulate objects using only visible light, Soft Matter 6 (15) (2010)3447e3449.
[10] M. Yamada, M. Kondo, R. Miyasato, Y. Naka, J.-i. Mamiya, M. Kinoshita,A. Shishido, Y. Yu, C.J. Barrett, T. Ikeda, Photomobile polymer materialsdva-rious three-dimensional movements, J. Mat. Chem. 19 (1) (2008) 60e62.
[11] M. Camacho-Lopez, H. Finkelmann, P. Palffy-Muhoray, M. Shelley, Fast liquid-crystal elastomer swims into the dark, Nat. Mater. 3 (5) (2004) 307e310.
[12] H. Finkelmann, E. Nishikawa, G. Pereira, M. Warner, A new opto-mechanicaleffect in solids, Phys. Rev. Lett. 87 (1) (2001) 015501.
[13] J. Cviklinski, A. Tajbakhsh, E.M. Terentjev, UV isomerisation in nematic elas-tomers as a route to photo-mechanical transducer, Eur. Phys. J. E 9 (1) (2002)427e434.
[14] K.D. Harris, R. Cuypers, P. Scheibe, C.L. van Oosten, C.W. Bastiaansen, J. Lub,D.J. Broer, Large amplitude light-induced motion in high elastic moduluspolymer actuators, J. Mater. Chem. 15 (47) (2005) 5043e5048.
[15] C. Van Oosten, K. Harris, C. Bastiaansen, D. Broer, Glassy photomechanicalliquid-crystal network actuators for microscale devices, Eur. Phys. J. E 23 (3)(2007) 329e336.
[16] Y. Yu, M. Nakano, A. Shishido, T. Shiono, T. Ikeda, Effect of cross-linkingdensity on photoinduced bending behavior of oriented liquid-crystallinenetwork films containing azobenzene, Chem. Mater. 16 (9) (2004)1637e1643.
[17] M. Kondo, M. Sugimoto, M. Yamada, Y. Naka, J.-i. Mamiya, M. Kinoshita,A. Shishido, Y. Yu, T. Ikeda, Effect of concentration of photoactive chromo-phores on photomechanical properties of crosslinked azobenzene liquid-crystalline polymers, J. Mater. Chem. 20 (1) (2010) 117e122.
[18] Y. Wu, Q. Zhang, A. Kanazawa, T. Shiono, T. Ikeda, Y. Nagase, Photoinducedalignment of polymer liquid crystals containing azobenzene moieties in theside chain. 5. Effect of the azo contents on alignment behavior and enhancedresponse, Macromolecules 32 (12) (1999) 3951e3956.
[19] K.M. Lee, H. Koerner, R.A. Vaia, T.J. Bunning, T.J. White, Relationship betweenthe photomechanical response and the thermomechanical properties ofazobenzene liquid crystalline polymer networks, Macromolecules 43 (19)(2010) 8185e8190.
[20] Y. Zhao, T. Ikeda, Smart Light-responsive Materials: Azobenzene-containingPolymers and Liquid Crystals, John Wiley & Sons, 2009.
[21] T. Fischer, L. L€asker, J. Stumpe, S. Kostromin, Photoinduced optical anisotropyin films of photochromic liquid crystalline polymers, J. Photochem. Photobiol.A Chem. 80 (1) (1994) 453e459.
[22] D. Corbett, M. Warner, Bleaching and stimulated recovery of dyes and ofphotocantilevers, Phys. Rev. E 77 (5) (2008) 051710.
[23] C.D. Eisenbach, Effect of polymer matrix on the cis-trans isomerization ofazobenzene residues in bulk polymers, Die Makromol. Chem. 179 (10) (1978)2489e2506.
[24] G.S. Kumar, D. Neckers, Photochemistry of azobenzene-containing polymers,Chem. Rev. 89 (8) (1989) 1915e1925.
[25] H.K. Kim, X.S. Wang, Y. Fujita, A. Sudo, H. Nishida, M. Fujii, T. Endo, Reversiblephoto-mechanical switching behavior of azobenzene-containing semi-inter-penetrating network under UV and visible light irradiation, Macromol. Chem.Phys. 206 (20) (2005) 2106e2111.
Adv. Funct. Mater. 15 (7) (2005) 1155e1159.[28] T.T. Mills, G.E. Toombes, S. Tristram-Nagle, D.-M. Smilgies, G.W. Feigenson,
J.F. Nagle, Order parameters and areas in fluid-phase oriented lipid mem-branes using wide angle X-ray scattering, Biophysical J. 95 (2) (2008)669e681.
[29] J. Seelig, W. Niederberger, Deuterium-labeled lipids as structural probes inliquid crystalline bilayers. Deuterium magnetic resonance study, J. Am. Chem.Soc. 96 (7) (1974) 2069e2072.
[30] K.M. Lee, N.V. Tabiryan, T.J. Bunning, T.J. White, Photomechanical mechanismand structure-property considerations in the generation of photomechanicalwork in glassy, azobenzene liquid crystal polymer networks, J. Mater. Chem.22 (2) (2012) 691e698.
[31] X. Tong, L. Cui, Y. Zhao, Confinement effects on photoalignment, photo-chemical phase transition, and thermochromic behavior of liquid crystallineazobenzene-containing diblock copolymers, Macromolecules 37 (9) (2004)3101e3112.
[32] S. Xiao, X. Lu, Q. Lu, B. Su, Photosensitive liquid-crystalline supramoleculesself-assembled from ionic liquid crystal and polyelectrolyte for laser-inducedoptical anisotropy, Macromolecules 41 (11) (2008) 3884e3892.
[34] C. Li, J.-H. Yun, H. Kim, M. Cho, Light propagation and photoactuation indensely cross-linked azobenzene-functionalized liquid-crystalline polymers:contribution of host and concerted isomerism, Macromolecules 49 (16) (2016)
6012e6020.[35] J.J. Wie, K.M. Lee, T.H. Ware, T.J. White, Twists and turns in glassy, liquid
crystalline polymer networks, Macromolecules 48 (4) (2015) 1087e1092.[36] R.W. Lenz, Synthesis and properties of thermotropic liquid crystal polymers
with main chain mesogenic units, Polym. J. 17 (1) (1985) 105e115.[37] C. Barrett, A. Natansohn, P. Rochon, Cis-trans thermal isomerization rates of
bound and doped azobenzenes in a series of polymers, Chem. Mater. 7 (5)(1995) 899e903.
[38] D.H. Wang, J.J. Wie, K.M. Lee, T.J. White, L.-S. Tan, Impact of backbone rigidityon the photomechanical response of glassy, azobenzene-functionalized pol-yimides, Macromolecules 47 (2) (2014) 659e667.
[39] C.S. Paik, H. Morawetz, Photochemical and thermal isomerization of azoar-omatic residues in the side chains and the backbone of polymers in bulk,Macromolecules 5 (2) (1972) 171e177.
[40] K. Anderle, R. Birenheide, M. Eich, J.H. Wendorff, Laser-induced reorientationof the optical axis in liquid-crystalline side chain polymers, Die Makromol.Chem. Rapid Commun. 10 (9) (1989) 477e483.
[41] Y. Wu, Y. Demachi, O. Tsutsumi, A. Kanazawa, T. Shiono, T. Ikeda, Photoin-duced alignment of polymer liquid crystals containing azobenzene moieties inthe side chain. 2. Effect of spacer length of the azobenzene unit on alignmentbehavior, Macromolecules 31 (4) (1998) 1104e1108.
[42] C.D. Eisenbach, Isomerization of aromatic azo chromophores in poly (ethylacrylate) networks and photomechanical effect, Polymer 21 (10) (1980)1175e1179.
